Synthesis, characterization and X-ray crystal structures of cyclam derivatives. 7. Hydrogen-bond induced allosteric effects and protonation cooperativity in a macrotricyclic bisdioxocyclam receptor

Article abstract of DOI:10.1039/b508076b

The unprecedented cooperative protonation properties displayed by a barrel-shaped macrotricyclic tetraamine incorporating two 14-membered bisamide rings maintained in a face-to-face arrangement is rationalized in terms of allosteric effects upon binding o

Full text of DOI:10.1039/b508076b

View Article Online / Journal Homepage / Table of Contents for this issue
L E T T E R
Synthesis, characterization and X-ray crystal structures of cyclam
derivatives. 7. Hydrogen-bond induced allosteric effects and
protonation cooperativity in a macrotricyclic bisdioxocyclam
receptorwz
´ ´
Michel Meyer, Laurent Fremond, Enrique Espinosa, Stephane Brandes, Guy Yves Vollmer
and Roger Guilard*
´
´
´
Laboratoire d’Ingenierie Moleculaire pour la Separation et les Applications des Gaz (LIMSAG,
´
´
CNRS UMR 5633), Universite de Bourgogne, Faculte des Sciences, 6 boulevard Gabriel,
21100 Dijon, France. E-mail: Roger.Guilard@u-bourgogne.fr; Fax: þ33 3 80 39 61 17;
Tel: 33 3 80 39 61 11
Received (in Montpellier, France) 7th June 2005, Accepted 29th June 2005
First published as an Advance Article on the web 19th July 2005
The unprecedented cooperative protonation properties displayed
by a barrel-shaped macrotricyclic tetraamine incorporating two
14-membered bisamide rings maintained in a face-to-face
arrangement is rationalized in terms of allosteric effects upon
binding of the first and third protons.
mass spectrum displayed only the [M þ H1] signal expected for
T1 (m/z ¼ 661) and T2 (m/z ¼ 605.5), demonstrating that no
oligomers were present. In CDCl3, two sets of sharp 1H NMR
signals, each corresponding to one half of the number of
hydrogen atoms, are observed at room temperature. Accord-
ingly, it is inferred that T1 exists as a mixture of two isomers in
slow exchange, a major (M) and a minor (m), for which the
orientation of both dioxocyclam units differs by a 1801 rotation
around the pivotal tertiary amines. Although a definitive
assignment is not possible, one form possesses a C2h and the
other a D2 time-averaged symmetry. With increasing tempera-
ture, all signals broaden and start to coalesce above 320 K,
indicating a fast tumbling of both rings, while cooling down to
210 K provokes a progressive disappearance of the resonances
arising from the m isomer. According to a van’t Hoff analysis
of the low-temperature spectra, the M # m interconversion
equilibrium is characterized by K298 ¼ 0.2, DH ¼ 10.2(1) kJ
Allosteric regulation is a well known and an important
possibility in biology to control the catalytic activity of en-
zymes,1 but it remains a rarity for synthetic systems.2 Increased
understanding of the fundamental principles behind supramo-
lecular chemistry has sparked growing interest in the design of
host molecules able to efficiently transmit information through
conformational changes mediated by the binding of a molecule
or an ion to a second remote coordination site.3 Association via
noncovalent interactions of allosteric effectors to artificial
receptors not only serve as means to alter the complexation
characteristics of the other sites, but furthermore provides a
new entry into the realm of molecular switches.4 In this regard,
knowledge of the factors that trigger proton-driven molecular
reorganization is of fundamental scientific interest.5
Herein, we report the unprecedented cooperative protona-
tion properties exhibited by macrotricyclic receptors of cylind-
rical topology possessing either two 5,12-dioxocyclam (T1) or
cyclam (T2) units maintained in a face-to-face arrangement
(Scheme 1). For the first time, we provide direct experimental
evidence (i.e. model free) that an abiotic, neutral tetraprotic
base is able to add two out of four protons cooperatively in
spite of positive charge accumulation.
molꢁ1, and DS ¼ 20.8(3) J molꢁ1 Kꢁ1
.
X-ray diffraction analysis of T1 ꢀ 4H2O provided helpful
insights about the conformation adopted by the free-base
ligand (Fig. 1). The central cavity, delimited by the phenyl
spacers on the lateral side, takes up a twisted arrangement with
a dihedral angle between both aromatic rings of 61.21. The
dioxocyclam subunits are rotated by 85.31 with respect to each
other so that the amide nitrogen atoms sit almost on the top of
the tertiary amines belonging to the facing ring. Most impor-
tantly, each amide proton interacts with the endo-oriented lone
pair of the adjacent tertiary amine, giving rise to two trans-
T1 was prepared in four steps with 26% yield starting from
5,12-dioxocyclam,6 while the octaamine T2 could be conveni-
ently isolated by quantitative reduction (BH3/THF) of T1.7
Alternative synthetic methods of T2 have also been described,8
but they do not allow the preparation of heteroditopic tricycles
incorporating two rings of different nature or size. En route
towards this goal, a more versatile procedure has been devised
which enabled T2 to be obtained in ca. 20% yield. The isolated
compounds, T1 ꢀ 4H2O and T2 ꢀ 8HCl ꢀ 6.5H2O were character-
ized by elemental analysis and TGA for their water content,
FTIR, 1H and 13C NMR spectroscopy. The MALDI/TOF
w For the previous paper in this series, see ref. 10.
z Electronic supplementary information (ESI) available: VT 1H NMR
spectra, 1H–1H NOESY chart, Curie plots, and 1H NMR titration of
T1; full experimental details. See http://dx.doi.org/10.1039/b508076b
Scheme 1
T h i s j o u r n a l i s & T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 5
N e w J . C h e m . , 2 0 0 5 , 2 9 , 1 1 2 1 – 1 1 2 4
1121

View Article Online
doublet of doublet centred at 9.25 ppm. Most intriguingly,
progressive acidification resulted in the vanishing of all reso-
nances arising from the free base and the simultaneous appear-
ance of a new set of peaks that is accompanied with severe line
broadening between p[H] 5.5–4.5. This behaviour is consistent
with slow proton transfers at the NMR time scale for the
partially protonated species, which seems to be a common
theme among dioxocyclam-based polycyclic receptors.10 Evi-
dence for nonhydrogen-bonded amide protons in the fully
protonated form is provided by the 0.53 ppm upfield shift
experienced by the NH signal. In contrast, the central xylyl
protons resonate at much lower field (Dd ¼ 0.66 ppm) upon
protonation, supporting their reorientation from a radial to a
tangential position with respect to the cylindrical cavity.
In order to unravel the structural and electronic conse-
quences of the b-alaninamide fragments on the basicity of
the neighboring amines, potentiometric titrations of tricycles
T1 and T2 and their monocyclic trans-dimethylated models (1
and 2) were carried out in the methanol–water mixed solvent.
Special care was taken to ensure that the equilibrium was
reached at each titration point. In a first approach, the multiple
binding events encountered by these polyprotic bases were
evaluated on hand of the occupancy concept and Scatchard
analysis.12 The site occupation factor r, which corresponds to
the average number of protons bound per molecule possessing
t chemically identical sites, was directly derived from the
experimental p[H] readings according to relation (1).
Fig. 1 ORTEP plot (thermal ellipsoids set at 50% probability) of T1
showing the trans-annular hydrogen bonds. Atoms labeled with primes
belong to the second half molecular unit, which is generated by a
twofold axis placed at (12, y, 43). Water molecules and all H(–C) hydrogen
atoms are omitted for clarity. Selection of interatomic distances and
angles: C17ꢀ ꢀ ꢀC170 ¼ 4.04(2) A, N1ꢀ ꢀ ꢀN40 ¼ 5.84(2) A, N2ꢀ ꢀ ꢀN30 ¼
6.35(2) A, N2ꢀ ꢀ ꢀN20 ¼ 6.29(2) A, N2ꢀ ꢀ ꢀN40 ¼ 6.39(2) A, N4ꢀ ꢀ ꢀN40 ¼
6.24(2) A, N2ꢀ ꢀ ꢀN4 ¼ 4.50(2) A, N1ꢀ ꢀ ꢀN2 ¼ 2.84(2) A, N1–H1ꢀ ꢀ ꢀN2 ¼
146.7(2)1, N3ꢀ ꢀ ꢀN4 ¼ 2.81(2) A, N3–H3ꢀ ꢀ ꢀN4 147.0(2)1.
annular hydrogen bonds spanning across each U-shaped
dioxocyclam ring [Hꢀ ꢀ ꢀN ¼ 2.03(2) and 2.07(2) A].
½H þꢂtot ꢁ 10ꢁp½Hꢂ
r ¼
with 0 ꢃ r ꢃ t
ð1Þ
Hydrogen bonding was also confirmed by the presence of a
broad nNH stretching mode at 3233 cmꢁ1 in the solid-state
½Lꢂtot
infrared spectrum of T1 ꢀ 4H2O. The diagnostic amide I (nCQ
O
t
P
¼ 1651 cmꢁ1) and amide II (dCNH ¼ 1554 cmꢁ1) absorption
bands appear in the expected range. Interestingly, the FTIR
spectrum recorded in diluted CDCl3 solutions were concentra-
tion independent and evidenced no significant shifts, empha-
sizing that the intramolecular N–Hꢀ ꢀ ꢀN hydrogen bonds are
preserved in solution. Furthermore, increasing the concentra-
tion of T1 did not lead to the appearance of a low-energy
satellite of the C–D stretching band arising from the solvent at
2253 cmꢁ1.9 This confirms that the tertiary amine’s lone pairs
are not exposed to the bulk, but rather adopt an endo-orienta-
tion as evidenced in the solid state.
i
ib01i½H þꢂ
i
Y
i¼1
r ¼
with b01i
¼
K01k
ð2Þ
t
P
i
1 þ b01i½H þꢂ
k¼1
i¼1
The Scatchard plots (r/[H1] vs. r) obtained for T1 and T2
(Fig. 2) clearly established that only four protons were taken
up in the p[H] range 11.9–2.0, although the biscyclam receptor
possesses eight protonation sites. Moreover, the hyperbolic
shape of the graph corresponding to T2 contrasted with the
downward-curved plot obtained for T1, highlighting a drasti-
Additional information on the solution structure and evi-
dence for intramolecular hydrogen bonding was provided by
1D and 2D NMR spectroscopy. Complete assignment of the
cross-peaks appearing in the 1H–1H NOESY correlation chart
further suggests that the main structural features of T1 seen in
the solid state are essentially preserved in CDCl3 (see ESIz).
The endo orientation of the amide protons (d ¼ 8.81 ppm) is
attested by NOE connectivities with the g-CH2 and the central
xylyl protons. Moreover, their engagement in intracyclic hy-
drogen bonds was confirmed by performing VT NMR experi-
ments. The slopes of the linear Curie plots obtained in pure
CDCl3 (Dd/DT ¼ ꢁ4.33 ppb Kꢁ1) but also in the non-deuter-
ated CH3OH/H2O (1 : 1 v/v) mixture (Dd/DT ¼ ꢁ5.16 ppb Kꢁ1
at pH 7) are consistent with solvent-protected N–H groups10
and fall in the typical range of temperature coefficient values
found for the yet stronger NHꢀ ꢀ ꢀO C interactions in peptides
Q
(Dd/DT o ꢁ5 ppb Kꢁ1 in CDCl3).11
The binary solvent mixture was mainly chosen for solubility
reasons. However, performing these experiments in an aqueous
non-deuterated medium offered the possibility to observe the
amide proton resonance, which proved to be an outstanding
probe for sensing the conformational rearrangements induced
by protonation of T1. In spite of an obscured aliphatic region,
spectral changes occurring in the aromatic part were conveni-
ently monitored as a function of HCl concentration (see ESIz).
At p[H] values (p[H] ¼ ꢁlog[H3O1]) where the free-base form
predominates, NH protons appear as a partially resolved
Fig. 2 Scatchard plots for (a) T1 and (b) T2. Experimental p[H] values
were converted into site occupation factors (r) using eqn. (1). Solid lines
were drawn according to eqn. (2) and the protonation constants
reported in Table 1. Solvent: CH3OH/H2O (1 : 1 v/v), I ¼ 0.1 (KCl),
T ¼ 298.2(1) K.
1122
N e w J . C h e m . , 2 0 0 5 , 2 9 , 1 1 2 1 – 1 1 2 4

View Article Online
cally different behavior of both ligands and providing direct
experimental evidence for cooperative protonation of T1.
Indeed, the shape of the plot is an excellent diagnostic tool
for detecting multiple-binding cooperativity, the phenomenon
by which the coordination of a given substrate enhances the
binding strength of the subsequent one. Deviation from linear-
ity (statistical binding) is a sufficient criterion to assess such a
positive (downward concavity) or negative (upward concavity)
effect.12 For statistical reasons, the ratio of the successive
equilibrium constants Ki11/Ki is less than unity for a polytopic
receptor having t identical and independent sites. Quantita-
tively, positive cooperativity is encountered when the experi-
mental Ki11/Ki ratio is larger than the critical factor for
statistical binding.
Once the unusual proton affinity of T1 has been established,
the stepwise protonation constants (Ki) were refined by
weighted nonlinear least-squares using the Hyperquad 2000
program.13 The accuracy of the calculated parameters reported
in Table 1 was ascertained by the excellent match between the
experimental Scatchard plot and the line derived from the
theoretical expression of r as a function of the equilibrium
constants.12 Moreover, Monte-Carlo simulations showed that
the shape of the calculated curve is highly sensitive to varia-
tions as small as 0.02 log units of the Ki values.
For T2, noncooperativity is evidenced by the regular de-
crease of nearly one logarithmic unit of the four stepwise
protonation constants, which is slightly higher than expected
for statistical binding. This sequence can be rationalized in
terms of charge accumulation and electrostatic repulsive
effects. Based on the facts that K1 and K2 are close to the
values measured for the model ligand 2, the first two protona-
tion steps of T2 involve the secondary amines, which are more
basic than the bridgehead tertiary nitrogen atoms. Accord-
ingly, proton binding is proposed to take place alternatively on
one cyclam subunit and then on the diametrically opposed site
belonging to the facing ring. Minimization of coulombic
repulsive interactions is best achieved by this sequence, classi-
cally observed for polyamines of various topologies.7,14,15
As expected, replacement of the secondary amines in 2 and
T2 by electron withdrawing amide groups confers a much
weaker basicity to the dioxocyclam derivatives 1 and T1.6
However, the intriguing inversion of both protonation con-
stants in the case of the model compound 1 is more surprising,
although protonation studies of 5,12-dioxocyclam also re-
vealed weak cooperativity.6 Since the relative magnitude of
K2 over K1 reflects the free-energy cost of protonating an amine
in the neighbourhood of a positively charged ammonium
group, it is concluded that the coulombic repulsion energy
for 1H221 is compensated by a favourable change in solvation
and/or strain energy.16
tically exacerbated when such fragments are built in a more
constrained tricyclic framework, as highlighted by the exotic
acid–base properties of T1. The strong cooperativity displayed
by this compound is reflected by a most unusual set of
protonation constants: K2 and K4 are respectively two and
one order of magnitude higher than K1 and K3. Moreover, the
two first and two last constants can be paired, while the second
equilibrium constants for 1 and T1 are virtually identical.
According to preliminary results obtained for structural ana-
logues of T1 that differ by the nature or length of both bridging
units, this trend seems to be a common feature. A literature
survey revealed only a very limited number of systems where
protonation enhances the basicity of the yet unprotonated
sites.15,17–19 Cooperativity has been attributed either to a
conformational rearrangement enforcing hydrogen bonding18
or by the encapsulation of mediating water molecules.17,20
Curiously, these compounds belong all to the cryptand family,
but none revealed such a strong cooperative effect as T1.
Both FTIR and 1H NMR results clearly support that the
free-base form of T1 retains approximately the solid-state
conformation in solution. Moreover, the unexpectedly low
K1 value found for T1 compared to the monocyclic analogue
1 strongly suggests that the four tertiary amines are still
engaged in trans-annular N–Hꢀ ꢀ ꢀNtert hydrogen bonds in
solution. Based on these experimental evidences, allosterism
might best account for the peculiar proton affinity. Considering
that the first entering proton will necessarily disrupt one of
those intracyclic interactions, a conformational rearrangement
of the protonated dioxocyclam ring is taking place, which in
turn will also break up the second trans-annular hydrogen
bond. Hence, the thereby disengaged electron doublet becomes
more accessible, thus increasing the basicity of the second
amine in the already monoprotonated dioxocyclam unit.16 This
proton-triggered allosteric effect accounts for the inversion of
K2 over K1 but also for the similar K2 values found for 1 and
the tricyclic cage T1. We conclude therefore that both amines
21
of the same dioxocyclam subunit are protonated in the T1H2
species and furthermore that the hydrogen-bond network is
preserved in the unprotonated ring (Scheme 2).
This allosteric mechanism can operate a second time as the
ligand takes up the third and fourth proton, in full agreement
with the experimentally observed order K4 4 K3. Due to
charge accumulation and electrostatic repulsion, the latter
values are slightly lower compared to K2 and K1, respectively.
It is worth noting that this phenomenon requires two weakly
mechanically-coupled subunits, as both must rearrange inde-
pendently from each other in order to avoid the simultaneous
disruption of all hydrogen bonds as the first amine protonates.
Flexibility of the tricyclic cage is therefore a prerequisite for
observing cooperative protonation in such systems. The ab-
normal proton and metal-binding properties induced by the
b-alaninamide moieties of T1 and the closely related dioxocy-
clen and dioxocyclam analogues will be reported in due course.
Noteworthy, the peculiar reversal of protonation constants
found for monocyclic 5,12-dioxocyclam derivatives is drama-
Table 1 Protonation constants of the various macrocyclesa
1
T1
4.62(7)
2
T2
log K1
log K2
log K3
log K4
6.36(1)
6.90(2)
10.41(6)
8.53(6)
10.13(1)
9.26(1)
7.69(2)
6.62(3)
0.87
6.73(8)
4.01(3)
5.27(4)
o2
o2
b
D1,2
D2,3
D3,4
ꢁ0.54
ꢁ2.11
1.88
b
2.72
ꢁ1.26
46.5
1.57
1.07
b
a
Solvent: CH3OH/H2O (1 : 1 v/v), I ¼ 0.1 (KCl), T ¼ 298.2(1) K. Ki
þ H1 # LHii1. Values in
(iꢁ1)1
refers to the equilibrium: LHiꢁ1
parenthesis correspond to the standard deviation in the last significant
b
digit. Di,j ¼ log Ki ꢁ log Kj. For statistical binding, the expected
values for four identical and independent sites are D1,2 ¼ D3,4
log 8/3 E 0.426 and D2,3 ¼ log 9/4 E 0.352.12
¼
Scheme 2
N e w J . C h e m . , 2 0 0 5 , 2 9 , 1 1 2 1 – 1 1 2 4
1123

View Article Online
Chem. Soc., 1997, 119, 4934–4944; (b) M. Meyer, A. M. Albrecht-
Gary, C. O. Dietrich-Buchecker and J. P. Sauvage, J. Am. Chem.
Soc., 1997, 119, 4599–4607.
V. W.-W. Yam, X.-X. Lu and C.-C. Ko, Angew. Chem. Int. Ed.,
2003, 42, 3385–3388.
I. Huc, Eur. J. Org. Chem., 2004, 17–29.
Experimental
All experimental details related to the synthesis, spectroscopic,
and potentiometric characterization of the described com-
pounds can be found in the Electronic Supplementary Infor-
mation (ESIz).
4
5
6
L. Fremond, E. Espinosa, M. Meyer, F. Denat, R. Guilard, V.
´
Huch and M. Veith, New J. Chem., 2000, 24, 959–966.
S. Brandes, F. Denat, S. Lacour, F. Rabiet, F. Barbette, P.
Pullumbi and R. Guilard, Eur. J. Org. Chem., 1998, 2349–2360.
(a) M. Lachkar, R. Guilard, A. Atmani, A. De Cian, J. Fischer
and R. Weiss, Inorg. Chem., 1998, 37, 1575–1584; (b) S. Develay,
R. Tripier, F. Chuburu, M. Le Baccon and H. Handel, Eur. J.
Org. Chem., 2003, 3047–3050.
7
8
.
Crystal data for T1 4H2O
A colourless crystal of prismatic shape (0.32 ꢄ 0.25 ꢄ 0.20 mm)
was obtained by slow evaporation from an ethanol/water (1 : 1
v/v) solution. C36H60N8O8, M ¼ 732.92, orthorhombic, a ¼
12.9970(2), b ¼ 18.4860(2), c ¼ 16.1270(3) A, U ¼ 3874.7(1)
A3, T ¼ 110(2) K, space group P bcn, Z ¼ 4, m(Mo-Ka) ¼
9
J. Cheney, J. P. Kintzinger and J. M. Lehn, Nouv. J. Chim., 1978,
2, 411–418.
10 M. Meyer, L. Fremond, A. Tabard, E. Espinosa, G. Y. Vollmer,
´
R. Guilard and Y. Dory, New J. Chem., 2005, 29, 99–108.
11 (a) H. Kessler, Angew. Chem., Int. Ed. Engl., 1982, 21, 512–523;
(b) T. Cierpicki and J. Otlewski, J. Biomol. NMR., 2001, 21,
249–261.
12 (a) B. Perlmutter-Hayman, Acc. Chem. Res., 1986, 19, 90–96; (b)
G. Ercolani, J. Am. Chem. Soc., 2003, 125, 16097–16103.
13 P. Gans, A. Sabatini and A. Vacca, Talanta, 1996, 43,
1739–1753.
14 (a) A. Damsyik, P. J. Davies, C. I. Keeble, M. R. Taylor and K. P.
Wainwright, J. Chem. Soc., Dalton Trans., 1998, 703–711;
(b) H. Plenio, C. Aberle, Y. A. Shihadeh, J. M. Lloris, R.
Martinez-Manez, T. Pardo and J. Soto, Chem.-Eur. J., 2001, 7,
2848–2861.
0.090 mmꢁ1, 10778 reflections measured, 5657 unique (Rint
¼
0.0630) which were used in all refinements on F2, 255 refined
parameters, 4 restrains, R1 ¼ 0.0532, wR2 ¼ 0.0992 [I 4 2s(I)],
R1 ¼ 0.1415, wR2 ¼ 0.1288 (all data). CCDC reference number
233399. See http://dx.doi.org/10.1039/b508076b for crystallo-
graphic data in CIF or other electronic format.
Acknowledgements
This work was supported by the Centre National de la
Recherche Scientifique (CNRS) and the Conseil Regional de
´
Bourgogne. L.F. and G.Y.V. are grateful to the Ministere de
l’Education Nationale, de la Recherche et de la Technologie
for a Ph.D. fellowship.
15 A. Bencini, A. Bianchi, E. Garcia-Espana, M. Micheloni and J. A.
Ramirez, Coord. Chem. Rev., 1999, 188, 97–156.
16 R.D. Hancock, R. J. Motekaitis, J. Mashishi, I. Cukrowski, J. H.
Reibenspies and A. E. Martell, J. Chem. Soc., Perkin Trans. 2,
1996, 1925–1929.
References
1
(a) J. Monod, J. P. Changeux and F. Jacob, J. Mol. Biol., 1963, 6,
306–329; (b) D. E. Koshland, in The Enzymes, ed. P. Boyer,
Academic Press, New York, 1970, vol. 1, p. 341–396.
17 (a) E. Graf, Ph.D thesis, Universite L. Pasteur, 1979; (b) B. Sarkar,
´
P. Mukhopadhyay and P. K. Bharadwaj, Coord. Chem. Rev.,
2003, 236, 1–13.
2
(a) J. Rebek, Acc. Chem. Res., 1984, 17, 258–264; (b) S. Shinkai,
M. Ikeda, A. Sugasaki and M. Takeuchi, Acc. Chem. Res., 2001,
34, 494–503; (c) M. Takeuchi, M. Ikeda, A. Sugasaki and S.
Shinkai, Acc. Chem. Res., 2001, 34, 865–873.
18 P. G. Potvin and M. H. Wong, Can. J. Chem., 1988, 66, 2914.
19 C. Bazzicalupi, P. Bandyopadhyay, A. Bencini, A. Bianchi, B.
Valtancoli, D. Bharadwaj, P. K. Bharadwaj and R. J. Butcher,
Eur. J. Inorg. Chem., 2000, 2111–2116.
3
(a) S. Blanc, P. Yakirevitch, E. Leize, M. Meyer, J. Libman, A.
Van Dorsselaer, A. M. Albrecht-Gary and A. Shanzer, J. Am.
20 D. K. Chand, K. G. Ragunathan, T. C. W. Mak and P. K.
Bharadwaj, J. Org. Chem., 1996, 61, 1169–1171.
1124
N e w J . C h e m . , 2 0 0 5 , 2 9 , 1 1 2 1 – 1 1 2 4